Many conventional CT scanner utilize X-ray radiation beams formed as one or more narrow fans, where each fan, with its corresponding line of detectors, defines a slice in the subject. The scan is performed either slice-by-slice by moving the subject along the scanner rotation axis and stopping it while a radial scan is taken, or a helical scan is performed where the subject moves simultaneously with the scanning.
Multi-slice scanning as described above with currently available scanners reaches a width in the order of 40 mm. Thus, to scan a typical organ requires translating the subject axially along the rotation axis. When moving organs are involved, however, especially the beating heart, the temporal resolution, i.e. ‘freezing’ the organ, is problematic. Various techniques were developed to tackle the problem such as gating by ECG or learning the rhythm, with varying degrees of success.
One area of current interest is CT scanning of the heart, referred to as “cardiac CT.” This presents especially difficult problems since the length of image that is required is of the order of 120 mm. Furthermore, for best results it is desirable to acquire all of the data in a one half plus fan beam angle rotation or in a single rotation. A half plus rotation can be made in 180 milliseconds, which, for proper phase within the diastole would result in sufficiently consistent data for reconstruction.
To overcome the limited width and reduce the numerous rotations of radiation/detector around the subject, some approaches were suggested where a multiplicity of radiation sources are used with linear or two-dimensional detectors. For example, U.S. Pat. Nos. 5,966,422, 7,016,455, 7,039,152 and 7,039,153 describe CT systems utilizing multiple fan beams sources arranged circumferentially about the CT rotation axis, each with a corresponding detector array. The disclosures of these references are incorporated herein by reference.
U.S. Pat. No. 5,625,661, the disclosure of which is incorporated by reference, describes a system in which a plurality of axially displaced sources of X-rays are used to irradiate the patient. This system has a number of rows of detectors equal to the number of sources. Since the number of sources would be limited and a required resolution would indicate a detector width of at most 2 mm, the system has a limited patient coverage and is described only in a spiral scan mode. Thus, this system does not allow for single or half rotation cardiac CT scanning.
U.S. Pat. Nos. 5,712,889 and 6,229,870, the disclosures of which are incorporated by reference, describe X-ray sources and systems in which plurality of axially displaced sources emit a plurality of parallel fan beams, formed by array of septa, each irradiating different parts of the scanned subject and directed at different part of the detector array. These systems are inherently inefficient in utilization of the X radiation since most radiation is absorbed by the septa while still requiring high power, generating high heat load, to energize the sources.
Cone-beam scanning is also known where a cone of radiation covers a substantial section of the subject. In cone beam scanning an X-ray source irradiates the subject over a relatively larger solid angle and an opposing area detector on the other side of the subject detects the transmitted radiation. Thus, a wide cone beam system covers a much larger portion of the subject relative to a fan beam or a narrow cone beam. With the larger area coverage, translating the subject is avoided or more limited, and a fast scanning is possible with, in principle, better temporal resolution.
The simplest cone beam scan mode is a rotation of a single source (and detector array) about the subject (circular source trajectory), although a partial scan (less than 360°) is also possible.
The reconstruction algorithm used for reconstructing cone beam acquired data is usually of the type called “Feldkamp” or “FDK” method (Feldkamp, L. A., L. Davis, and J. Kress (1984). Practical Cone-beam Algorithm. Journal of the Optical Society of America 1, 612-619, the disclosure of which is incorporated herein by reference) or a derivative thereof. The FDK algorithm is approximate and allows the scanned subject to extend outside the projected region in the axial direction. Unfortunately, with these types of solutions the reconstructed images have artifacts that worsen with the distance from the mid plane (the source trajectory plane) due to “data incompleteness”. Similar problems occur with other reconstruction algorithms known in the art.
It is usually accepted that data completeness leading to exact reconstruction is available if the Tuy-Smith condition is met (Tuy, H. (1983). An inversion formula for cone-beam reconstruction. SIAM Journal of Applied Mathematics 43, 546-552, the disclosure of which is incorporated herein by reference). This condition requires that the source trajectory shall intersect every plane passing through the scanned subject volume of interest. This is definitely not the case for a cone beam single circular source trajectory.
FIG. 1 shows a simplified schematic axial cross-sectional presentation of the use of a single cone beam to irradiate a patient. The use of very large cone angles, sufficient to encompass the heart have been mooted and are under development. However, as can be understood from the following discussion, they are expensive and unnecessarily irradiate the patient.
FIG. 1 shows a single cone beam system 10 of the prior art having a single source 12 and a detector array 14. It should be understood that each detector element 16 shown in FIG. 1 corresponds to a linear circumferential or planar linear array of detector elements. In a typical geometry each detector elements covers 0.5 mm width at an axis of rotation 58. Reference 18 represents a cylinder of reconstruction (patient). Cone beam 20 is seen to intersect patient 18 in three distinct regions. A first region 22 is a region in which data can be acquired over all rotation angles and is referred to herein as the region of full coverage. A second region 24 on each side of region 22 corresponds to regions of data acquisition, in which data is available over all rotation angles only at diameters smaller than the scanned subject diameter. Region 26 on either side of region 24 corresponds to volume irradiated only at particular gantry angles, in which data is not available for a half or a whole rotation. Since the region of full coverage is so much smaller than the spread of the beam at the detector for imaging the heart in a single rotation, the number of detector rows must be increased from typically 240 rows to cover the heart to about 280 rows. Further, larger parts of the patient are subject to ionizing radiation than should optimally be irradiated. With the geometry of FIG. 1, even in region 24 the data does not meet the Tuy-Smith condition, except for data close the focal spot rotation plan, resulting in degraded image quality for parts of the subject away from the rotation plan.
U.S. Pat. No. 6,996,204, the disclosure of which is incorporated herein by reference describe a method in which a cone beam scanner with geometry similar to that described in FIG. 1 is used in steps: cone beam attenuation data is acquired for one circular trajectory of the source, the scanned subject is axially translated relative to the source, cone beam data is acquired again for a second circular trajectory of the source such that the two cone beams have a volume of overlap and images are reconstructed where for volume elements in the volume of overlap data is used from both source trajectories. The method disclosed reduces the amount of missing data and improves the resulted image quality. However, the method cannot be used for a single heart beat cardiac scanning since it involves two acquisition cycles separated in time.
U.S. Pat. Nos. 5,068,882 and 5,187,659, the disclosures of which are incorporated herein by reference describe systems wherein two overlapping cone beam sources are provided and data is acquired from both within a single rotation of the sources relative to the scanned subject. In the disclosed embodiments, the sources are displaced both axially and radially and each is provided with a two dimensional detector array, thus increasing cost and complexity of the systems.
U.S. Pat. Nos. 7,072,436 and 7,145,981 describe systems in which two dimensional arrays of radiation sources are provided with extent in both axial and angular directions. A common detector array is provided to receive attenuated radiation from said sources.
The use of X-ray beams in medical imaging is well established. X-ray projection imaging is used to obtain static or fluoroscopic images. Computerized tomography (CT) scanning uses a plurality of X-ray images to assemble a 3D reconstruction of an organ or a potion of a subject. In both cases, single focal spots are typically used. However, there are imaging systems using multiple focal spots.
U.S. Pat. No. 6,181,771 to Hell describes electro magnetic deflection of an electron beam emanating from a cathode in an X-ray tube using a pair of electro-magnets placed on opposite sides of the beam. Hell is concerned primarily with focusing the electron beam at a focal spot on an anode. The disclosure of this patent is fully incorporated herein by reference.
U.S. Pat. No. 6,483,890 to Malamud, describes an X-ray tube with a single cathode and a single anode. The X-ray tubes described by Malamud includes four deflection plates arranged in pairs. Malamud describes using a changing voltage applied between each pair to cause an electron beam to rotate in a circle about its nominal axial path through a midpoint between the four deflection plates. Malamud is concerned primarily with focusing the beam at different places on the anode using a predetermined temporal pattern. The disclosure of this patent is fully incorporated herein by reference.
U.S. Pat. No. 4,689,809 to Sohval describes an X-ray tube with a similar physical configuration to that of Malamud and/or Hell. Sohval, like Malamud and Hell uses an electromagnetic field to deflect a beam of electrons originating from a cathode in an X-ray tube. The disclosure of this patent is fully incorporated herein by reference. U.S. Pat. No. 4,637,040 describes a CT system utilizing this X-ray tube.
U.S. Pat. No. 4,912,739 to Weiss also describes use of an electromagnetic field to deflect a beam of electrons originating from a cathode in an X-ray tube. The disclosure of this patent is fully incorporated herein by reference.
U.S. Pat. No. 6,229,870 to Morgan describes a plurality of discrete anodes mounted within the vacuum envelope, the anodes selectively generating a plurality of parallel X-ray beams. According to Morgan, each anode element is associated with a cathode assembly selectably excitable by a filament power supply. When selected, each cathode assembly generates an electron stream which strikes the corresponding anode element and produces X-ray beams. The disclosure of this patent is fully incorporated herein by reference.
Grid control and Grid pulsing for X-ray tubes is well known and has been commercialized, for example in the MRC line of X-ray tubes of Phillips Medical Systems.
US Patent Publication No. 2006/0285633 to Sukovic, the contents of which are incorporated by reference herein, describes a multi-source CT scanner with a plurality of cone-beam X-ray sources offset along a CT axis. The CT scanner includes a controller programmed to alternately activate a first and second X-ray source. It is further stated that although the X-ray sources in both embodiments described are shown as completely separate, a single, multiple-source unit could also be utilized where a power supply or other circuitry are shared between the multiple sources.
US Patent Publication No. 2005/0195935 to Yahata, the contents of which is incorporated by reference herein, describes an X-ray generator that generates two cone beam X-radiations which pass through two successive slabs. It is described that an X-ray tube permitting alternate X-irradiations includes a double set of an anode, a cathode, a grid and a switch. One of a 0 bias voltage and a negative bias voltage is applied to each of the grids via the switch. The negative voltage assumes a value permitting blocking of an electron flow between a cathode and an anode. Consequently, X-irradiation can be discontinued by changing the connections to be made via the switches. The connections to be made via the switch are changed with either of signals that are out of phase with each other. It is mentioned that since the X-ray tube has two grids associated with the two pairs of an anode and a cathode, the timings of generating two cone beam X-radiations can be controlled.